The concept layer.
Defined once.

Limits: Does not prescribe universe membership, revision handling, volume allocation, EMA initialization, or software behavior.

Limits: Does not define the pilot's exhaustive partition, causal cutoff, revision lineage, EMA policy, or normalized breadth formula.

Limits: Does not prescribe one universal initialization, missing-session, revision-resolution, rounding, or software interface policy.

Limits: Interpretive descriptions and historical conventions are not current out-of-sample performance evidence or a universal fresh-series calibration.

Limits: Historical interpretive thresholds are not universal trading rules; the source does not prescribe the package's evidence or software failure policies.

Limits: The article's market interpretation is excluded as evidence of forecasting, causation, thresholds, or profitable trading behavior.

Limits: Historical publication conventions do not define a universal modern volume feed, adjustment basis, revision policy, or predictive result.

Limits: Platform symbols, displayed thresholds, universe choices, and implementation details are vendor-specific rather than universal methodology requirements.

Limits: Defines governed platform behavior rather than claiming these policies are universally prescribed by external McClellan methodology.

Limits: Product-specific and licensed; it does not create universal eligibility, repair, imputation, tolerance, or aggregation rules.

Limits: A serialization profile does not define financial event, effective, publication, availability, or processing semantics.

Limits: The database does not contain exchange sessions, trading holidays, or product calendars.

Limits: Hours vary by venue, product, date, rule version, and exceptional notice; they are not a universal calendar.

Limits: The normative text is licensed; an ISIN does not by itself identify an issuer, venue, listing, or eternal internal entity.

Limits: Identifier hierarchy does not establish legal rights, fungibility, or universal persistence, and provider terms still apply.

Limits: A MIC identifies a market or segment, not a security, issuer, listing, or trading symbol.

Limits: Current directories are not complete historical identity chains, and symbols are not globally unique permanent identifiers.

Limits: Implementations use negotiated profiles; FIX fields do not impose one universal feed, stale-data, deduplication, or reconciliation policy.

Limits: It supplies a provenance vocabulary, not a financial-market lineage schema or retention policy.

Limits: It defines instance validation, not financial semantics, release severity, compatibility, precision, or schema-drift policy.

Limits: Authoritative for the provider's indices; conventions are not universal defaults for every index or security-level return series.

Limits: Provider-specific rules and discretion must be read with the methodology effective for the named index and date.

Limits: Rules are specific to applicable FTSE Russell indices and can contain market, eligibility, tax, and timing exceptions.

Limits: General education does not define every event field, tax result, exchange rule, data-provider adjustment, or index treatment.

Limits: CRSP conventions, fields, codes, and licensed observations are provider-specific and are not reproduced as universal rules.

Limits: Principles are not an index calculation formula, a substitute for product methodology, or a complete statement of applicable law.

Limits: MSCI methodology and production data are provider-specific and branded data or calculated values may require licenses.

Limits: The library links multiple documents with separate versions and applicability; the effective document must be identified.

Limits: A delisting filing does not establish final shareholder proceeds, terminal value, successor consideration, or reconstructed return.

Limits: Product coverage, field semantics, timing, history, access, and redistribution are specific to Nasdaq and its delivery products.

Limits: Provider methodology and production constraints are specific to MSCI indices and do not define one universal optimizer.

Limits: Authoritative for the named index family; it does not create a universal volatility-control formula or investment result.

Limits: Later index constraints, covariance estimators, investability rules, and production optimization controls are separate choices.

Limits: Covariance estimation, constraints, convergence, turnover, and index-provider implementation remain governed choices.

Limits: Metric selection, accounting definitions, normalization, data lags, rebalancing, and production rules must be declared.

Limits: Production rates, holidays, currencies, tenors, and calculation details require the full applicable methodology and data licenses.

Limits: EDGAR access guidance does not define a research system's local ingestion, validation, revision-selection, or availability boundary.

Limits: Terminology does not prove local availability, select a research revision, or interpret the accounting effect of an amendment.

Limits: The page documents a platform indicator; universe membership and data handling must still be declared by the implementation.

Limits: Examples are educational and provider-specific; they do not prescribe a universal universe, lookback, or trading rule.

Limits: Threshold interpretation and universe composition are contextual and must not be treated as guaranteed signals.

Limits: Provider symbols and eligibility rules may differ from another data vendor's point-in-time universe.

Limits: Interpretive levels are market- and period-dependent; zero denominators and feed coverage require explicit policy.

Limits: The starting level is arbitrary and different exchange feeds can produce different histories.

Limits: The glossary definition does not define a complete production data contract or denominator-zero policy.

Limits: Historical signal descriptions are not evidence of future performance, and implementations must freeze thresholds and window semantics.

Limits: Antitrust concentration thresholds do not automatically apply to index-constituent concentration; this glossary reuses only the mathematics.

Limits: Applying the general diffusion construction to sectors or factor groups is a declared Fintech Builder adaptation.

Limits: Display colors and pattern interpretations are conventions rather than universal predictive rules.

Limits: The short glossary entry does not define bar aggregation, data corrections, thresholds, or pattern-detection rules.

Limits: The contract is specific to Binance Spot streams and does not define every venue's candle or finalization semantics.

Limits: TA-Lib defaults are library conventions and do not establish universal definitions for small or large candle bodies.

Limits: The source does not prescribe a financial candle-classification window or threshold factor.

Limits: Forecasting notation and initialization can differ from technical-indicator conventions and must be translated explicitly.

Limits: The reference does not choose the newest-heavy linear weights used by one technical-analysis WMA convention.

Limits: The source is an RSI implementation; applying the recurrence to a standalone RMA requires an explicit input and readiness contract.

Limits: Provider lookback and unstable-period behavior must be distinguished from another implementation's warm-up policy.

Limits: The documentation does not make TEMA equivalent to the unrelated TRIX oscillator or guarantee reduced error.

Limits: Rounding non-integer lengths and warm-up behavior require explicit implementation choices.

Limits: Implementation seeding, readiness, and parameter defaults must be frozen for reproducible parity.

Limits: The abstract alone is insufficient for code parity; exact coefficients, state, clamps, and warm-up require a pinned implementation contract.

Limits: Platform defaults, initialization, and display conventions are not universal definitions or performance evidence.

Limits: Provider-specific compatibility behavior and unstable periods must be reconciled before claiming parity.

Limits: Moving-average type, denominator-zero handling, signal line, and histogram must be frozen by the consuming methodology.

Limits: Exact tie handling, window endpoints, and implementation readiness require a complete pinned calculation contract.

Limits: The pinned version supports reproducibility but does not make every platform's tie convention equivalent.

Limits: The bibliographic page alone is not an executable specification; equality, seeds, and state transitions require pinned rules.

Limits: Zero-denominator and compatibility behavior can differ from the Fintech Builder package and must be stated explicitly.

Limits: Unstable periods and platform-specific first-ready indexes must be reconciled before claiming exact parity.

Limits: Default periods and plotting offsets are conventions; calculation origin must remain separate from display coordinate.

Limits: Initial direction and extended-SAR options differ across implementations and must be declared.

Limits: ATR smoothing, initialization, equality rules, and sign conventions can vary by platform.

Limits: Seed, smoothing convention, lookback, and missing-bar handling must be declared for reproducible use.

Limits: Seed, flat-series behavior, unstable periods, and platform compatibility must be frozen before claiming parity.

Limits: Moving-average types, zero-range policy, and first-ready indexes can differ across implementations.

Limits: Output scale, nested warm-up, RSI seed, smoothing type, and flat-RSI-window behavior require explicit policy.

Limits: Lookback endpoints, zero-range behavior, and threshold interpretation remain implementation and research choices.

Limits: The common 0.015 constant is a convention, and zero-deviation or mixed-adjustment inputs require explicit handling.

Limits: Window ordering, weights, first-ready index, and zero summed-range behavior must be declared.

Limits: Long/short ordering, EMA seeds, warm-up suppression, optional signal line, and zero denominator vary by implementation.

Limits: The moving branch can change; exact parity requires a pinned revision and an explicit tie and ranking-window contract.

Limits: CCI specifically uses the mean of absolute deviations of typical prices from their window arithmetic mean; other centers are different measures.

Limits: TA-Lib leaves the first row unavailable; the local D07 contract explicitly uses a high-low fallback.

Limits: Library control flow does not prescribe the local diagnostic driver label or its author-defined tie priority.

Limits: TA-Lib readiness follows its own unavailable-first-TR convention and therefore differs from the local alignment.